Author: Albert Silver
Date: 18:12:27 02/14/00
Go up one level in this thread
On February 14, 2000 at 01:11:07, Vincent Vega wrote:
>On February 13, 2000 at 17:46:16, Albert Silver wrote:
>
>>So he is studying the Elo value of each ply for only one program? What's the
>>point? He'll have to analyze the value for each and every ply along the line,
>>and even then the results will only be for that one program.
>
>The point is to find out whether there is a falloff in elo gain when plies
>increase. Why it is important should be obvious if you followed any related
>discussion - if such a falloff exists, it means that for this program (I believe
>Crafty was to be used) even exponential processor speed increases (Moore’s law)
>will have a diminishing impact on ratings. The point of my original post was
>that I believe that if such a falloff exists (previous data says it does not),
>high-knowledge programs (and programs with high-branching factors) will have an
>edge in the future because they will be hurt less.
>
>>Why is it no-no, and why do you have to use time to get anything significant? If
>>you want to refute what I've said about it, I'd like to at least see the
>>reasoning.
>
>It is a no-no because 1. programs can easily have different branching factors so
>one more ply for one program may mean 4 times as many nodes evaluated and 7
>times as many for another, 2. to get to the 10th ply for one program may mean
>that it has evaluated 10,000 positions for 0.01 second each, for another
>10,000,000,000 positions for 0.1 second each, and for yet another 10,000,000
>positions for 0.0001 second each, 3. evaluation of that fixed ply could be
>similarly skewed. As a result, fixing plies at some level, as you tried to do,
>doesn't lead to any significant results because it handicaps programs with fast
>evaluations and low branching factors for no reason and makes inter-program
>comparisons invalid.
A little miscommunication problem here. I wasn't trying to compare different
programs, one with a large eval function, and another with none, as of course it
would be an exercise in futility for the very reasons you stated above. Instead,
have the same program (with a reasonably sized eval) twice, one version with the
eval all de-activated except the material values, and the other working as
usual. Now play them at a fixed depth.
>
>If you think you can restate your arguments using time, I'd like to hear them.
>I'm afraid that using plies to compare different programs is fatally flawed.
No, time is actually what is futile IMO. Time can only yield one thing: greater
depth in plies and everything that that ply can bring to the program. What else
can time bring if not depth? Let's try this again: take a program and reduce it
so that it knows the laws of chess plus the material values only. What can it
learn with each and every ply? It can either calculate a forced material win or
a mate, but if it finds neither, it's eval will always remain zero. Always. It
will never learn anything about the position. It will never display a half pawn
advantage, nor an edge for anyside unless it sees a forced material win. In
fact, unless it finds either a material win or a mate, it's results will be
simply list of the moves that don't lead to either for it and a flipping of the
coin to see which gets played. It can do this at 10 plies, 20 plies, or 30
plies, and unless something is forced, it will know NOTHING. It's eval will be a
big fat +0.00.
>Time is an obvious constant because there is no way for a program to decide to
>use more time and still play another program (I guess it could use less, but for
>all the chess programs I know of, that would be counterproductive).
>
>>How much higher? In my opinion this probability is close to nil unless the
>>position has already been compromised; yet how will Mindblack achieve this great
>>position with random move choices?
>
>You are making an error here. The playing strength of Mindblank and CyberGM
>isn't in question here. The question is how much they gain by multiplying the
>time they have to evaluate the position. So if after 10 mins, Mindblank
>evaluation was weak, the question is how less-weak will it be after 20 mins, not
See above. It will be the same.
>whether it will be as good as that of CyberGM. If it's better by an average of
>100 points than it was, but CyberGM's is better by 150 points than it was, that
>means CyberGM gains more (I'm not saying whether it does or not, that's just an
>example).
>
>>I imagine that writing an ultrafast program with no knowledge would be extremely
>>easy. Try it. I think you'll find that it will lose so fast it won't even be
>>funny. Do you think it will magically start coordinating an attack against the
>>enemy king without any knowledge?
>
>That's the same error as before. Of course it will be bad. I never doubted
>that. That's not the issue, though.
>
>>I suspect, no offense, that although you know how to play, that you aren't very
>>strong. I say this because I tend to see these arguments that knowledge is
>>over-rated by people who do not realize just how deep knowledge can go. A GM has
>>_skills_ that a 2200 player simply doesn't have. It goes beyond the simple 'he
>>plays better positional moves and calculates deeper'.
>
>What in heavens made you think that I believe knowledge is overrated? That's a
>complete opposite of what I stated in several posts already. I think knowledge
>is the way to go and I like for CSTAL to do well exactly because it's said to be
>knowledge-based (even if its programmer is somewhat disliked here). I don't
>know if it will be able to match current leaders, but I hope it will. Please
>answer that question so I’ll know where my statements may have been unclear.
You are right, you said that about CS-Tal in the beginning, and then you also
said that Mindblank, which only knows how to count pawns and pieces, will
somehow improve its evaluation with time (or greater depth as that is what time
brings). These two statements are in contradiction hence the lack of clarity,
though I can see this was involuntary.
>
>Even when I was just starting playing chess it was absolutely clear to me that
>better players not only saw deeper but also had superior knowledge of the game.
>Was it different for you? Did you only see tactics when you were starting out
>and you didn't notice any greater understanding of the game in superior players?
> If not why would you assume that a potentially weaker player would make the
>same mistake?
Yes, and no. Chess is kind of funny in that each horizon that one reaches, shows
us a little more of what we don't know, and therefore a new horizon to go after
is revealed. Most players don't get that far along the scale and therefore never
realize that this process just keeps on going. Most players I've encountered
seem to believe that the difference between a 2200 player and a GM is only that
the GM is doing it better and deeper. Your comment on Mindblank learning with
increased depth is what led me to believe that you believed the same.
Albert Silver
>
>BTW, when I was in my teens I played competitively for a while until I found
>other interests. I did quite well back then. Now my playing strength has
>probably declined quite a bit, I don't really know.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.